既可以使用数组实列的方法，也可以调用顶层的 Numpy 函数。

In [92]: arr = np.random.randn(5, 4)

In [93]: arr
Out[93]:
array([[ 0.40078213,  0.18221019,  0.33039728, -0.17005462],
       [ 0.06459178, -1.25292774,  0.22437965,  0.30838516],
       [-0.40874228, -0.68919994,  1.65989984,  0.73510201],
       [-0.07695719, -0.25494756,  0.45742598, -1.2483983 ],
       [ 0.01360322,  1.41888667,  1.43997912,  1.1402712 ]])

In [94]: arr.mean()
Out[94]: 0.21373432899920713

In [95]: np.mean(arr)
Out[95]: 0.21373432899920713

In [96]: arr.sum()
Out[96]: 4.2746865799841425

In [97]: arr.mean(axis=1)
Out[97]: array([ 0.18583375, -0.16389279,  0.3242649 , -0.28071927,  1.00318505])

In [98]: arr.sum(axis=0)
Out[98]: array([-0.00672234, -0.59597839,  4.11208186,  0.76530544])

arr.mean(1)表示每一列的平均值，arr.sum(0)表示行和。

In [99]: arr = np.array([0, 1, 2, 3, 4, 5, 6, 7])

In [100]: arr.cumsum()
Out[100]: array([ 0,  1,  3,  6, 10, 15, 21, 28], dtype=int32)
In [102]: arr = np.array([[0, 1, 2], [3, 4, 5], [6, 7, 8]])

In [103]: arr
Out[103]:
array([[0, 1, 2],
       [3, 4, 5],
       [6, 7, 8]])

In [104]: arr.cumsum(axis = 0)
Out[104]:
array([[ 0,  1,  2],
       [ 3,  5,  7],
       [ 9, 12, 15]], dtype=int32)

In [105]: arr.cumprod(axis = 1)
Out[105]:
array([[  0,   0,   0],
       [  3,  12,  60],
       [  6,  42, 336]], dtype=int32)

0代表列的计算，1代表行的计算，即对列和行分别累积求和、积。（《利用python进行数据分析》113页）